Multicolour Imaging: Moving From Theory to Practice

Multicolour fluorescence imaging, when rigorously applied, allows the selective visualization of two or more features within a sample. The qualification “when rigorously applied” is so important because multicolour imaging is prone to artifacts. For multicolour experiments, separate are used to image fluorescent labels with different fluorescence and emission spectra. Problems can arise, however, when the fluorescence emission from one fluorophore spills over into the detection channels used to image the other fluorophores labelling the sample.

Match filters and fluorophores

To avoid artifacts, the first step is to inspect whether the fluorophore spectra are well-matched to the filters available. Each filter set should excite and detect one fluorophore at a time without contamination from the other fluorophores in the sample. To illustrate, let’s consider a direct immunofluorescence imaging experiment. Say we want to test whether two specific proteins, Protein A and Protein B, are localized in the same vesicles within a cell. We plan to label the proteins with primary antibodies. One antibody, conjugated to the fluorophore Lumi490, targets Protein A, while the second antibody is conjugated to Lumi550 and targets Protein B (note: Lumi490 and Lumi550 are hypothetical examples and both antibodies are raised in different species). 

Filter selection example

We plan to image the immunofluorescence labelling on a wide-field microscope, where each spectral channel corresponds to a separate filter cube, consisting of an excitation, dichroic, and emission filter. Figure 1 depicts the excitation spectra of Lumi490 and Lumi550, along with the corresponding fluorescence excitation filters available. Likewise, Figure 2 plots the fluorescence emission spectra and the corresponding emission filters. After checking for possible overlap in both the excitation and emission, we note that the exciting Lumi490 may also excite Lumi550 as the fluorescence excitation spectra shows some overlap with the excitation wavelengths (the blue excitation filter, used to excite Lumi490, may also excite some fluorescence from Lumi550). This may create issues when imaging the Lumi490-conjugated antibody because the signal may be contaminated by the fluorescence emission from Lumi550. After inspecting the fluorescence emission spectra and associated filters, however, we predict that very little fluorescence emission from Lumi550 will be detected when Lumi490 is imaged because the emission filter used shows very little overlap with the fluorescence emission spectrum of Lumi550. We conclude that your fluorophore choice is suitable, and proceed to label your samples. Do you agree with this assessment?

After labelling your samples, you acquire images on a wide-field microscope equipped with filter cubes. You record images sequentially, that is, one channel at a time. When you compare the images acquired with each channel, you note that they look similar, and in fact, show perfect overlap within the resolution limit of your microscope. What should you do next?

• Choice 1: You have done due diligence by checking the spectra for possible overlap. You can see that the signal overlaps perfectly in both channels, indicating that they are indeed localized to the same compartment in the cell. You are ready to start building on this result.

• Choice 2: Consider and test for false positives.

Although the first choice may seem like a reasonable and efficient choice, the correct choice is the second one.  It is important to prepare and image single-labelled controls, even if the spectra pass the initial checks in the planning stage because fluorescence spectra can be misleading. Spectra are typically plotted using arbitrary units. One of the fluorophores may be much brighter than the other, and what may look like minor overlap may translate to bright contaminating signal at the microscope. In summary, checking that the spectra match fluorescence filters is the starting point, but it must be validated by experimental checks at the microscope. 

Verify filter and fluorophore selection

To carry out these checks, it is important to image the dual-labelled sample, and then to use identical acquisition settings to image the single-labelled sample using both channels. In this way, you can test whether the detected fluorescence signal is emitted from the intended target, with minimal contamination from the other labelled targets in the sample. It is also advisable to image an unlabeled control because sometimes autofluorescent signal can be confused with the specifically labelled features in the sample. 

Yet if you detect spectral contamination between channels, don’t despair- it does not mean you have to start over from scratch. If the detected spillover between channels is low, you can use quantitative approaches to correct for spectral overlap between channels. Further, other techniques, including spectral and/or lifetime imaging also offer alternate approaches to separating overlapping fluorescence emission. What’s key is that you recognize that fluorescence signal may not always arise from the intended target and that checking and correcting for spillover among the channels is an important part of any multi-colour imaging experiment. 

Have you ever had to deal with similar issues when imaging using multiple channels? Do you have any favorite resources that cover best practices for multicolour imaging? Feel free to share your responses here. I have also listed some of my favorites below.

Recommended resources

An excellent short video guide to the fluorescence microscope: https://youtu.be/PE3XBYebtZQ
Clear and succinct introduction to multicolour imaging: https://youtu.be/D_vxHMWC2mQ
A thorough introduction to colocalization imagsing: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3074624/
Considerations specific for superresolution microscopy: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4501387/